VCO: AM Hartley LO
1.2Mhz-2.2Mhz Hartley
@2012/5/7
* Output: 2Vpp
* Supply 9V-12V
* Hartley topology
* Revised version: Output 5Vpp (@6/10/2012)
STEP 1 try qrp.pops.net VFO schematic : @2012/5/7
Fig Sch.1
(NOTE: C2 should be removed)
this sch base on http://www.qrp.pops.net/images/2011/VFO-2011/oct12-SG.png.
My first version inductor of the Hartley oscillator is an HELICAL COIL wound on 1cm diameter plastic tube, length about 3.5cm. 80T tap at 10T, start oscillation from 1Mhz to 4Mhz. all circuit built into a minty box.
problems :
transform L1' second wing got very bad waveform, especially frequncy> 3Mhz, exactly like this photo.I suspected it's cause by minty box, box might had bad effect when metal wall close to the Helical coil.
But this kind of distortion very likely to be the non-linear load. the red enamelled wire in Fig coil.1 is the primary, the secondary is loosely coupled(via air). this might be insufficient to driven the cap of oscilloscope's probe. but after the buffer , output from the FT37-43's secondary still present bad waveform.
finally, the magnetic couple from L1 to Ft37-43 is the root cause. put a FT37-43 transformer in the minty, we get 1Vpp@3Mhz signal via magnetic couple, move it 15 cm away from L1, signal pickup is unnoticeable.
solution:
1) use shield oscillator inductor: @2012/5/8, use shield L1 is better
2) shield FT37-43 transformer if possible
3) use small, Ferrite core coil for Hartley Oscillator
another radiation source:
the long wire from 360p variable capacitor to the board, that's very strong radiation.
STEP 2 : trade-off and reconfigure @5/17 2012
we must make trade of about how much power the LO should output. for BJT type mixer a typical LO level is 200 mVpp, an double-balance Diode mixer might need 7 dBm, even 10 dBm , 20 dBm it depend on your goal of IIP3 and the Max RF input power( LO at least +10 dB compared to RF input). an Single FET(3-6 dB gain) might need 2 Vpp LO, cascade FET(10-15 dB gain) mixer might need 5 Vpp(~20 dBm) for best performance.
I want a LO with buffer Amplifier get a 10 dBm ( 2Vpp @50R) output, this will enable it's could be use to testing each type of mixer i had. the Fig Sch.1's buffer amplifier is attempt to keep the output amplitude stable instead of high level output.
To get the maximum conversion gain the LO node should be a short circuit at the RF and IF frequencies, so a 50R output is very necessary. (while the RF node should be a short circuit at the LO frequency to prevent the LO leakage into the RF port.)
here is another popular buffer Amp recommended by <<ARRL hand book>>, and it's performance is proven:
Fig buffer.2 output power about 10dBm-- 12dBm (2Vpp-- 2.5Vpp)
it accully give an output power only around 1Vpp.
for the NFB buffer, T1 given 20t:5t=4:1 voltage convert, so output 2Vpp needed from T1's primary signal reach to 8Vpp, NFB gain is (27k)/10k~=2.7(simulation result is 2.5 times voltage gain at Q3's collector), so Q1 source signal should strong to 8V/2.7 ~= 3Vpp, total tank voltage need to be >10Vpp. i checked my circuits, the output from L1's secondary of Fig Sch.1 is 1.4Vpp( builded circuits T1 ratio (Fig buffer.2 ) is 26:7. )
My hartley oscillator use a transformer with 50t:7t=7:1, total tank voltage is 10Vpp. the Hartley couple output voltage is less the expectation.
Possible fix:
1) adjust NFB Amp gain
2) use another type JFE for hartley, I use MPF102, most of them pinch off voltage is less than 4V would limited the tank voltage, if use J310 tank voltage will get better, compensate the the light couple Hartley output.
3) rewind the Oscillator transformer.
STEP 3. get more NFB Amp Gain and bias consideration
Fig. original
Q2,Q3 : 2sc9018
T1 as a inductive load, make possible max signal swing at T1's primary about 2*VCC= 20Vpp (inductive load make Q3 collect voltage can exceed the VCC, like a current source). Q2,Q3 form a NFB amplifier, the voltage gain(Q3's collector) is R2/R3, at the same time R2:R3 define the bias(10k could be DC isolated), this is the limitation to get high gain: assume 10mA for quiescent current, R8 eat 1.8V, Q2 base at lest bias at 3.2V, so high gain require high VCC: 3.2*2.7.
first step, replace R3 with 4.7k resistor, get gain about 6.7, if don't touch any other part of the NFB amp, we get a signal like this:
this is typical wrong bias result, we need 3.2V*6.7=22V VCC, that's too high, a typical receiver use 9V-12V supply. to low the supply requirement, remove R8, C5, connect Q3's emitter direct to the ground, then everything is ok, Q2 base still need around 1.8V to correct bias. need a supply: 1.8V*6.7=12V. experiment also confirm this, left image show the correct bias VCC>12V, right is @<11V:
Final schematic
Fig. 2Vpp
Hartley T0: 50t:7t
FT32-23: 26t:7t
Tank voltage: 5Vpeak
Hartley output: 0.7Vpeak
Q3 Collector: 4Vpeak
Output level: 1Vpeak
total current: 25mA
NOTE:
1) high gain degrade the performance a little bit
2) without R8 degrade the bias stability
STEP 4. double the Gain again? lower the supply?
A) double the gain
replace R3 as 2.7k, gain 11, output swing 2Vpeak, needed power supply 1.8V*11= 19.8V(plus R4 voltage down, we need around 23V), and quiescent current reach 50mA total.
B) low the supply
insert D4 to X, Y help to reduce supply requirement, but introduce harmonics . left one insert one Diode, right one insert 2 diode.
2 diode for bias damage the signal, one is acceptable. one diode get around 3V off from supply requirement, depend on the gain.
C) other good way might help a lot
the main problem is feed back resistor R3 also provide bias for Q2,Q3. to decouple them is the best method with more flexible and few more component.
D) more bias problem syndrome (bias current sufficient, Reduce R8 in Fig. original or step up supply)
Revised version
@6/10 2012
for high level output, now output 3Vpp. suitable driven a JFET mixer(which pinch off voltage around 3V):Mixer-J0
Test configure: Antenna-C0, Mixer-J0, LO-H0
output level: 3Vpp@1Mhz, 3.5V@2Mhz
MW AM radio listen experience:
1) noise background is lower if Mixer-J0's output transformer T1 is tune away from IF stage's band.
seems IF stage is lack of gain but out door test get more stations than Fig. 2Vpp. [10 stations from 700kzh to 1.54Mhz, LO-H0
not cover all MW band]
2) R6=> 36R for stabilization and reduce heat of Q3. and problem described in 1) might because the IF stage @6/18/2012,